Classifying Players For Unique Game Experiences 167
togelius writes "Whenever you play a game of Tomb Raider: Underworld, heaps of data about your playing style is collected at Eidos' servers. Researchers at the Center for Computer Games Research have now mined this data to identify the different types of player behavior (PDF). Using self-organizing neural networks, they classified players as either Veterans, Solvers, Pacifists or Runners. It turns out people play the game for very different reasons and focus on different parts of the game, but almost everyone falls into one of these categories. These neural networks can now quickly determine which of these groups you belong to based on just seeing you play. In the near future, such networks will be used to adapt games like Tomb Raider while they are played (e.g. by removing or adding puzzles and enemies), so you get the game you want."
The four types (Score:5, Informative)
In case anyone else was trying to figure out these roles... (page 6 last two paragraphs - > page 7)
Veterans = The power gamers, deaths usually only environmental.
Solvers = Die often (mainly from falling), methodical, slow.
Pacifists = Cannon fodder basically.
Runners = They run, they die, they run. The first thing that comes to mind here is a player that goes for the flag immediately in CTF.
Re:Almost everyone? (Score:4, Informative)
Re:Almost everyone? (Score:5, Informative)
The system discovers the categories. The analysis finds groupings of players who behave in similar ways through the game, and the researchers named those after-the-fact. There's no a priori reason why the players should group at all, though - the study could've equally found that only a small percentage of players clustered and the majority were radically different from each other.
Re:Bartle did this work already (Score:4, Informative)
Here [mud.co.uk]. But you could have found that yourself on Wikipedia...
Re:The four types (Score:4, Informative)
8.6% of players were Veterans, "players that die very few times; their death is caused mainly by the environment and they complete TRU very fast."
22.12% of players were Solvers. "Their long completion times, low number of deaths by enemies or environment effects indicate a slow-moving, careful style of play with the number one cause of death being falling (jumping).
46.18% of players were Pacifists: "The total number of their deaths varies a lot but their completion times are below average and their help requests are minimal indicating a certain amount of skill at playing the game.
16.56% of players were Runners, "players that die quite often and mainly by opponents and the environment. These players are very fast in completing the game (similar to the Veterans), while having a varying number of help requests which cover the majority of the H value range."
Re:Almost everyone? (Score:3, Informative)
Maybe it's someone who just like to run straight into a crowd of enemies and immediately die repeatedly for hours on end.
Runner
Or someone who logs in and just sits there not moving for hours. Or any number of other things that probably wouldn't even be classified as "play"
Pacifist or Solver
See, that's you doing it backwards -- once the categories have been defined by the neural networks (and labeled after the fact by humans), you are now trying to take any given data point and fit it into one of the categories. That's not how it works. Imagine looking at a 2D image containing many dots; if you were asked to draw perimeters around any significant clusters, you could probably do so without difficulty -- but depending on the 2D image you are given, it is entirely possible (even probable) that not every dot is going to be part of a cluster. If you wanted to include every dot, you could instead subdivide the image into regions, but that is a different task.
To speak to the specific examples, if Runner has been defined as something akin to "someone who likes to achieve the game's goals in as short a period as possible, skipping past subgoals and rewards if they present significant risk or slowdown", then in that case "someone who just like to run straight into a crowd of enemies and immediately die repeatedly for hours on end" could not be defined as a Runner. Similarly, if a Pacifist is defined as "someone who attempts to accomplish the game's goals using the minimum amount of violence possible", and Solver is defined as "someone who attempts to accomplish every task presented in the most efficient way possible, even when such tasks are not necessary to progress in the game", then "someone who logs in and just sits there not moving for hours" could not be defined as a Solver or a Pacifist. If you want to assign one of the labels to any given player description, sure you can do that, but it doesn't mean that player is actually part of the data cluster which defines the category.
Re:Thanks for the heads up (Score:3, Informative)
I think they are, though I'm sure that they're a little overwhelmed by the amount of data involved.
In the most recent patch 3.2 they removed "twinks" from regular battlegrounds and added XP. The vast majority of us cheered.